CHAPTER Democracy and the Politics 11 of Poverty Reduction

283

11CHAPTER

Democracy and the Politics of Poverty Reduction

The types of development strategies states pursue and the possibilities for achieving a redistribution of income, wealth and social benefi ts depend substantially on poli-tics. Politics refers to processes of cooperation, confl ict and negotiation that shape decisions about the production, dis-tribution and use of resources. Outcomes from the political process further depend on the way power is confi gured or distributed, the types of relationships governments estab-lish with different groups in society, and the institutions that structure relations and mediate confl icts among com-peting interests. Organization, contestation and claims-making by groups with strong ties to the poor are crucial in producing redistributive outcomes that reduce poverty. However, interest groups or social movements concerned with the welfare of the poor rarely organize around issues of poverty per se. Instead, they frame their discourse around rights, asset distribution, services, or earnings and benefi ts derived from work. This opens up the possibility of address-ing the structural roots of poverty, social rights and issues related to redistribution.

Even though poverty has been eradicated under certain types of authoritarian rule, the aim of this chapter is to convey an understanding of how it can be eliminated within a demo-cratic context, given that democracy, rights, freedom and choice are essential components of development.1 Although democracies offer opportunities for participation and contes-tation in policy making, redistributive outcomes cannot be taken for granted, given differences in the capacity of groups to organize, contest and infl uence public policy in different contexts. The challenge lies in forging democratic states that are both developmental and socially inclusive.

Current democracies face two types of constraints. The fi rst, underscored vividly by the fi nancial crisis of 2008–2009, is the capture of economic policy by investors, fi nancial insti-tutions and donors through various types of conditionality.

This has tended to favour stabilization and liberalization policies that are detrimental to welfare-enhancing struc-tural change. The second constraint relates to the limited nature of industrial transformation in most new democra-cies, the uneven quality of their democratic institutions and processes, and ethnic cleavages in many that shape choices and capacity for collective action. All of these factors affect the formation and growth of interest groups, social move-ments and parties necessary for holding leaders accountable and for constructing pacts with redistributive agendas.

Although democracies offer opportunities for participation and contestation in policy making, redistributive outcomes cannot be taken for granted

Although the World Bank and IMF–led Poverty Reduction Strategy Papers (PRSPs) support the participation of civil society groups in programme design, the process has mostly been reduced to consultation rather than participation to effect real change. The type of participation associated with the social pacts that produced rapid poverty reduction in the past differs substantially from the bargaining regime of the PRSPs, which relies heavily on non-governmental organizations (NGOs), and where the balance of power is strongly weighted against non-governmental groups.

This chapter shows that democracies have been able to deliver outcomes that are benefi cial to the poor under certain conditions – when:

rights are institutionalized, which allows the • poor to exercise political choice, build alliances with others and hold leaders to account;

COMBATING POVERTY AND INEQUALITY

284

groups with strong ties to the poor develop • capacity for independent organization and mobilization, transcend or reconcile horizontal divisions, and establish structural links with actors involved in policy making, leading, at times, to social pacts. Some success can be achieved without formal group ties to state actors, but this usually requires high levels of contestation and continuous mobilization to sustain gains; andelections are competitive and there is a • high probability that governments may lose offi ce, which can serve as an incentive for redistribution. However, electoral competitiveness without effective group organization and contestation may produce weak redistributive outcomes.

The poor suffer when interest groups and social movements are weak and the electoral system is not suffi ciently com-petitive. Technocratic styles of policy making – in which macroeconomic policy is insulated from public scrutiny – and aid conditionality also limit the scope for contestation and effective participation of civic groups with redistribu-tive agendas in the policy process.

Although growth and redistribution may at times appear to be in confl ict, they have been reconciled in industrialized democracies when employment levels and rates of unioni-zation are high, allowing unions to project an encompass-ing or national stance in bargaining that transcends narrow union interest. In welfare democracies with lower levels of industrialization, redistributive politics may also be encom-passing or inclusive when groups forge rural-urban alliances that incorporate wide segments of the working poor. Rural-urban alliances make it possible to extend welfare rights to all categories of citizens. In such democracies, leading par-ties tend to embrace a discourse of social rights, and groups engage the state as rights-bearing citizens rather than as clients. Fiercely contested elections, sometimes leading to changes in government, encourage parties to retain their social movement characteristics, which are essential for sustaining proactive links to the poor.

In sum, the chapter draws three main conclusions:Power relations are at the core of • development. Strategies that aim to reduce poverty and inequality must also consider ways to tip the balance of power.Although democracy tends to produce • redistributive outcomes, such outcomes should not be taken for granted. Group activism of various kinds – including voting power, bargaining and direct action – is important in generating shifts in power and ensuring that governments respond to social needs.Reducing poverty requires an expansion of the • bargaining power of the poor and those who represent them. This entails action on several fronts: rejecting technocratic styles of decision making, which limit the involvement of citizens and their representatives in policy-making processes; widening the participation of civil society in the policy process; institutionalizing civil and political rights to encourage broader participation in policy making; and supporting organizations that seek to protect the rights of both formal and informal workers.

Section 1 of this chapter highlights recent trends in democ-ratization and conceptualizes the links between democracy and redistribution, including the roles of interest groups and social movements in the politics of redistribution.

Section 2 discusses the constraints of technocratic styles of governance and aid conditionality on policy space.

Section 3 examines three types of domestic constraints to pur-suing redistributive policies: limited industrialization, the une-ven quality of democratic institutions, and ethnic diversity.

Section 4 analyses the conditions under which democra-cies deliver redistributive outcomes by examining fi ve cases involving activism by interest groups and social movements.

Section 5 draws conclusions and spells out implications for policy.

SECTION THREE – CHAPTER 11 – DEMOCRACY AND THE POLITICS OF POVERTY REDUCTION

285

1. Democracy and Redistribution

The 1980s and 1990s witnessed a strong wave of democ-ratization that affected most regions of the world. This was driven in part by societal pressures to force govern-ments to correct the failed policies of growth and redis-tribution associated with structural adjustment (in Africa and Latin America) and ineffi cient central planning (in Central and Eastern Europe), or to improve politi-cal and economic choices as economies matured (in East Asia). Added pressure came from multilateral and bilateral aid agencies, which expected that democratization would compel governments to formulate and implement what they perceived to be better economic policies and cut down on corruption.

Over the past 25 years, countries across all categories of national income have held competitive elections

There is a good deal of disagreement about the number of countries that can be classifi ed as democratic, although all datasets point to a sharp increase in the last 25 years. Polity IV reports an increase from about 40 democratic countries in 1980 to more than 90 by 2006, and a corresponding decrease in authoritarian regimes from more than 80 to about 25 over the same period. According to Freedom House, over 60 per cent of coun-tries could be classifi ed as democratic by 2000. Figure 11.1 indicates an increase in the number of low- and middle-income countries that held elections from 1975–2006 and in those in which the winner received less than 75 per cent of seats in the lower chamber. Interestingly, since the mid-1990s, the number of countries holding competitive elections has been evenly distributed among all categories of national income.

FIGURE 11.1: Competitive elections in low- and middle-income countries

0

20

40

60

80

100

197

5

Num

ber o

f cou

ntrie

sLow-income Lower-middle Upper-middle

197

619

77

197

819

79

198

019

81

198

219

83

198

419

85

198

619

87

198

819

89

199

019

91

199

219

93

199

419

95

199

619

97

199

819

99

20

00

20

01

20

02

20

03

20

04

20

05

20

06

Source: Sheingate 2008, drawing on Beck et al. (2001).

Democracy tends to produce redistributive

outcomes, but is subject to group infl uences

Democracy is defi ned as a system of government in which leaders periodically renew their mandates through free, fair and competitive elections. In addition, it is a system that acknowledges a set of rights – such as those of expression, organization and collective action – that allow citizens to exercise political choice and hold leaders accountable. Despite the centrality of elections in democracies, this chapter argues that redistribution requires additional mech-anisms of interest representation in the policy process.

A positive association between democracy and redistribution is posited by the median-voter theorem. According to the theorem, under universal suffrage, the median voter will earn the median income. However, when income is unequally dis-tributed, the median income falls below the mean income. Redistribution is expected to follow democratization because the mean income in pre-democratic societies is universally higher than the income of the median voter. Since the deci-sive voter earns a below-average income, it is assumed that he or she will favour a higher tax rate and redistributive


286

policies. In sum, democracy brings more people with below-average incomes to the polls, and they collectively press the government to redistribute income downwards.2

This way of conceptualizing democracy and redistribution is devoid of group infl uences, since it assumes that all vot-ers earning incomes below the median are a homogenous group that will automatically vote for redistribution. How-ever, behaviour is shaped by other factors, including rela-tionships in the workplace, by social and cultural settings, and by the political environment. Voting itself is a collec-tive action mediated by organized groups with compet-ing preferences and requires concerted effort for effective outcomes. It needs to be combined with other modes of organization to affect redistributive outcomes.

In the nineteenth and early twentieth centuries, the exten-sion of the vote to non-propertied groups was strongly associ-ated with redistribution. Democratization was contested by the rich, who saw it as a process that would facilitate tax increases by parliaments controlled by the rising working-class majority.3 Indeed, studies suggest that social spending, social programmes and taxation in major European countries increased and inequalities declined after countries embarked on democratization.4 These redistributive outcomes and sub-sequent expansion of welfare were, however, based on spe-cifi c types of group action, as represented by union activism, cross-class collective action, the strength of Leftist parties and their participation in government.5 How individuals organize into groups and affect public policy is therefore important in understanding the politics of redistribution.

The organization of groupsGroup organization may take three broad forms: as social movements, as interest-based associations and networks, and as political parties. Social movements emerge when protests against specifi c issues are linked to other efforts to address similar concerns and are sustained over time and across different locations. They are often diffuse and may not always be formally organized. Interest-based associations, which may emerge from, or support, social movements, include trade unions, farmers associations, professional and business organizations, neighbourhood groups, women’s

organizations and advocacy groups. The politics of produc-tion-based interest associations have, in some contexts, created a bargaining regime that places them fi rmly within the institutions of policy making rather than as lobbyists operating from the outside. Groups may combine voting power, bargaining and direct action to improve welfare.

How individuals organize into groups and affect public policy is important in understanding the politics of redistribution

Relatively few social movements or associations emerge specifi cally around the issue of poverty.6 Rather, they tend to focus on issues such as rights, asset protection or redis-tribution, access to services, or work-related earnings and benefi ts. However, the fact that movements emerge around issues that drive poverty and address them through politi-cal action means that they have the effect of politicizing poverty, placing it in its broader relational context. It also means that in certain contexts, the processes that can serve to deepen poverty might also serve to create the demand for movements contesting these same processes. Social movements may require interest-based associations to keep them active and mobilized.

In advanced industrialized democracies, movements, asso-ciations and parties tended to cluster according to the basic capital-labour cleavage associated with industrialization.7 Indeed, many political parties and interest associations were a product of social movements.8 In developing country con-texts, the three forms of organization do not cluster according to the basic industrial cleavage, although social movements and interest-based associations have provided a foundational base for parties that have embraced redistributive policies.

Current democracies face certain external and domestic constraints in pursuing redistributive policies. The external constraints have two dimensions. The fi rst involves pressure by investors and fi nancial institutions to narrow economic


287

policy making to a limited set of market-enhancing objec-tives, which elected governments and organized groups fi nd diffi cult to challenge even when they participate in defi n-ing poverty reduction strategies. The second relates to the power of bilateral donors in the policy-making systems of aid-dependent democracies exercised through various types of conditionality. Both constraints limit policy space for alternatives. The domestic constraints relate to the nature of industrial transformation and restructuring, the uneven quality of institutions and processes, and ethnic divisions. The next two sections describe these constraints before discussing group activism for redistribution.

2. External Constraints to Pursuing Redistributive Policies

Welfare states were formed when governments enjoyed rela-tive autonomy in using fi scal and monetary instruments to compensate losers for external shocks and facilitate redistribu-tion as economies grew. However, current patterns of globali-zation exert pressure for cuts in social expenditures and taxes and for more restrictive monetary policies. Firms in tradable sectors often pressure governments to lower taxes and reduce budget defi cits, and they perceive social programmes as addi-tional costs that will undermine international competitive-ness. Similarly, capital mobility gives fi rms, especially those in the fi nancial sector, the freedom to shift funds around or pun-ish governments that pursue fi scal policies that are inconsist-ent with business interests. In some countries, the structural change associated with liberalization has weakened fi rms that enjoyed a good deal of protection under import-substituting industrialization, including the social actors and welfare enti-tlements that underpinned such a regime.9

Government capture by technocrats insulates

decisions from the will of the people

Although the actual effects of globalization on welfare spend-ing and outcomes are still being debated,10 governments

are under considerable pressure to limit policy options to a narrow set of objectives that emphasize fi scal restraint, privatization and liberalization. In order to meet these objectives, governments often prefer to limit policy mak-ing to technocrats, or those with expertise, whose deci-sions are insulated from political processes. This, in effect, renders parliaments and social groups ineffective players in the policy-making process. Technocrats are mostly found in central banks and fi nance and trade ministries, which are integrated into global fi nancial and trade systems and increasingly adopt a less developmental and redistributive stance in policy making. They have gained strength at the expense of social and production-sector ministries.

Technocratic styles of policy making distort structures of accountability since governments tend to answer more to multilateral agencies and investors than to representa-tive institutions and the wider public. This affects the way governments respond to issues of employment and social policies, which may be sidelined or forced to conform to predetermined policy objectives that emphasize fi scal sta-bility. In democracies, legislative institutions are expected to articulate the choices of citizens, scrutinize government policies and provide legitimacy for policy outcomes. But economic policies affect social groups and institutions differently, and democratic processes and accountabil-ity suffer when important decisions about trade-offs are entrusted exclusively to technocrats. Central bank chiefs, for instance, whose institutions now enjoy a good deal of autonomy in new democracies, may be beholden to spe-cial-interest groups in the fi nancial world. This may give added privilege to strategies for reducing infl ation, as well as fi nancial and trade liberalization, over those of generat-ing employment or more inclusive social policies.

UNRISD research11 suggests that countries that are dependent on multilateral fi nancial institutions show high levels of policy capture by these institutions. Those with a longer history of democracy in which policy making refl ects compromises between politicians and citizens have pursued more heterodox policies. In some cases, the rise of technocratic behaviour in the executive branch can serve as an incentive for legislators to become more technically


288

competent. This may check the growth of technocracy if it is accompanied by improved levels of accountability to citizens, and if the latter also become more aware of the technical issues involved in policy making.

Relations between the executive and legislative branches in the economic policy fi eld in low-income democracies are often lopsided, a problem that is compounded by the lack of expertise of most legislators. For instance, more than 60 per cent of members in Malawi’s fi rst democratically elected par-liament had qualifi cations below a college diploma, and only about 15 per cent had a fi rst degree or more. While donor interventions have strengthened the technical knowledge of the executive branch, those of the legislature remain severely underdeveloped, making it diffi cult for lawmakers to properly represent the interests of their constituents and hold the executive branch to account. In many aid-depend-ent countries, multilateral fi nancial institutions have played important roles in identifying, supporting and recruiting technocrats for vital economic institutions. In Malawi dur-ing the 1990s, tensions existed between the government’s views on poverty alleviation and the demands of its struc-tural adjustment programme (SAP). The poverty allevia-tion programme launched in 1995 was never implemented because of the subsequent introduction of cash budgeting, the need to control expenditures, and the transfer of the technical staff entrusted with implementing the poverty alleviation programme from the National Economic Council to the Ministry of Finance and Economic Planning.

If a democratic transition is preceded by an authoritar-ian regime with success in delivering some economic outcomes, or one that has imposed restrictions on radical constitutional change, as in Chile in the 1990s, the new government may be forced to accept the former regime’s policy-making parameters, limiting the extent to which redistributive policies can be pursued. Elite consensus on economic policies may also emerge if past approaches to policy making are discredited, and parties and voters demand a new beginning. In the Czech Republic and Hun-gary, widespread distrust of central planning, along with strong faith in the market and membership in the Euro-pean Union as vehicles for reversing economic decline,

served to bring elites together across the political spectrum during the transition period. Parliamentary oversight of the executive branch mainly entailed acquiring informa-tion about the branch’s activities rather than contesting its policy choices. In Hungary, the consensus broke down in 1995 following a hard-hitting stabilization programme that led to party splits and intense parliamentary debates on the budget. In the Czech Republic, the crisis in 1996–1998 led to government criticism of the central bank’s monetary pol-icy. When the government later collapsed, it was replaced by a caretaker administration headed by the central bank governor. Parliamentary oversight of the executive branch improved when a new parliament was constituted in which no party had a majority.

Strengthening the accountability of parliaments Parliamentary accountability can be enhanced when ruling parties engage in critical support of government policies – rather than rubber-stamping them – and opposition parties develop expertise in bargaining with the executive branch and dispense with strategies of non-cooperation. A techno-cratically inclined executive cannot be held accountable if the legislative branch lacks comparable levels of technical expertise. And if parties fail to strengthen members’ tech-nical capacity, their parliaments will remain equally unin-terested in doing so. Until the early 1990s, very few efforts were made by the president of Argentina to consult parlia-ment, and a record number of decrees were used to push through tough reforms that affected welfare adversely. The rise of technocracy in the executive branch coincided with a decline in the operational budget, staff size and compe-tence of the parliament. Unable to scrutinize the executive on the basis of the technical merits of policy, opposition parties in parliament adopted strategies of non-cooperation. However, the technical knowledge and expertise of parlia-ment improved between 1995 and 1997,12 when it emulated the executive branch by incorporating more legislators who had been trained as economists. Subsequently, there was more willingness to question the executive, introduce bills and modify government-proposed legislation. In Chile, the number of economist-legislators has grown dramatically since the 1990s, and by 2000 they dominated the fi nance committees of both the Senate and Chamber of Deputies.


289

These economists have played an important role in moder-ating executive dominance in economic policy. However, parliament still lacks suffi cient expertise, information and resources to scrutinize all economic policies.

One of the issues that has been raised with respect to the inability of women legislators (brought into parliament through quotas) to defend women’s issues in public policies

relates to their frequent lack of technical expertise, espe-cially with respect to economic and fi scal policies. However, efforts are being made by some women parliamentarians (in South Africa, for example) to use their expertise in gender budgeting to strengthen their work. Box 11.1 discusses the extent to which women in India and South Africa have been able to use their presence in parliament to bring the issue of gender equality into policy making.

BOX 11.1: Translating women’s political representation into outcomes that promote equality

Indicators of women’s political empowerment have improved considerably since the advent of democratization. Starting in the

1970s, the creation of gender machineries fostered the idea of women as a constituency for policy makers to consider. More recently,

deliberate strategies such as quotas and reserved seats have facilitated women’s political access to parliaments. South Africa, for

example, has adopted national and local quotas, leading to signifi cant increases in women’s political representation. India has resisted

national quotas, but does have a system of reserved seats that is being implemented at the local level of panchayats. At the national

level, female representation remains below 8 per cent despite 60 years of democracy. At the local level, the introduction of reserved

seats triggered dramatic results, with nearly one million women entering representative institutions in the fi rst round of panchayat

elections (1994 and 1995).

Yet, while women have gained signifi cant access to state bureaucracies and legislatures, institutional access has not automatically

translated into political voice, and voice has not led to redistributive policy outcomes. Whether or not women’s inclusion in representative

institutions leads to public policy that enhances equality depends on the overall political context. Three factors warrant mention.

• The design of institutions has a bearing on whether citizens are able to exercise control over spending of public resources.

Low levels of citizen control reduce incentives for participation and limit the potential for pro-poor decision making.

• Competitive electoral systems are more likely to create opportunities for women’s organizations to insert their claims on processes

of public deliberation by leveraging differences among political parties. Similarly, strongly institutionalized parties with a high

degree of legitimacy can be pre-eminent vehicles for policy infl uence. Where parties are closed systems of patronage, the inclusion

of women on party lists will have little effect on policy outcomes.

• Civil society plays a crucial role in ensuring legitimacy for women political representatives and holding elected leaders accountable.

The presence and autonomy of women’s organizations is thus an important variable for equality outcomes. Especially where women’s

organizations see their interventions as going beyond inclusion to encompass accountability, their impact on those institutions is

likely to be considerable. If they are successful in building accountability and shifting spending patterns, women’s organizations can

build trust in formal institutions. Without such trust, it is unlikely that poor women will expend much energy in articulating their

interests through the state and that quotas will kick-start long-term sustained participation.

In South Africa, the combination of a single-party-dominant political system, limited devolution of fi nancial decision making and a

weakly autonomous civil society has limited the impact of the large number of women who have been elected to political offi ce since

1994. In India, by contrast, stronger decentralization and deeper histories of independent organizations of women and poor people

have created political space for newly elected women to re-shape spending priorities.

Source: Hassim 2009.


290

Aid conditionality limits the policy

options of governments

A second, related type of external constraint applies mainly to aid-dependent democracies.13 Such democracies are exposed to various types of conditionality and high lev-els of donor involvement in policy making, which limits the space of national actors to shape development agen-das. The most well-known, and criticized, is that of policy conditionality, which multilateral donors deployed to push through the SAPs of the 1980s. Under policy condition-ality, recipients commit to certain policies before aid is granted. Problems of national ownership and slippage in implementation encouraged donors to turn to other aid modalities, though not necessarily to abandon policy con-ditionality altogether.

Aid conditionality and high levels of donor involvement in policy making limit the space of national actors to shape development agendas

Process conditionalityProcess conditionality has gained infl uence as an alterna-tive aid modality, along with democratization and donor-fi nanced poverty reduction strategies. Under this type of conditionality, the policy-making process, rather than a particular policy, is the subject of donor concern. It seeks to change the balance of interests in policy making so that the general population, especially groups that repre-sent the poor, have a greater say. Process conditionality is evident both in direct interactions between donors and recipients, as well as in the PRSP process. It can be seen as an effort to democratize economic policy making. How-ever, the outcomes in most countries suggest that donor infl uence has greatly increased through process condition-ality, and the space for policy initiatives by national actors has narrowed.

The PRSPs call for the participation of civil society groups in their preparation, on the assumption that such partici-pation will result in pro-poor policies. Considerable debate has subsequently been focused on who is included in the participatory process and their credentials as representa-tives of the poor. In practice, this role has usually been taken on by NGOs, rather than associations of informal and formal workers, farmers or artisans, whose livelihoods are likely to be directly affected by the macroeconomic policies favoured by donors. Research in many countries, in fact, has found that important issues are often left out of discussions. One common omission has been that of mac-roeconomic frameworks, which are largely based on the International Monetary Fund’s (IMF) Poverty Reduction and Growth Facility and negotiated between governments and the IMF. As a result, the macroeconomic framework that constitutes the foundation of a country’s development strategy is subject to little or no debate. Moreover, though countries are required to involve civil society groups in the preparation of the PRSPs, there are no specifi c requirements as to the form that involvement should take.14 Participants’ recommendations often do not affect the fi nal form of the PRSPs, which depend on the approval of the international fi nancial institutions (IFIs) for funding. In many cases, civil society involvement is limited to taking part in con-sultations, rather than playing a more meaningful role in infl uencing the actual content of policies. As described in more detail in section 4, this form of participation departs fundamentally from the bargaining regime his-torically associated with improvements in the welfare of subaltern groups.

In many cases, civil society involvement in PRSPs is limited to taking part in consultations, rather than playing a more meaningful role in infl uencing the content of policies


291

Outcome-based conditionalitiesAnother type of conditionality that is gaining infl uence, especially among bilateral donors, is aid selectivity. This calls for aid allocation to be based on demonstrated suc-cess in policy implementation or achievement of develop-ment goals. The World Bank’s Assessing Aid report15 has been extremely infl uential in this regard. It concludes that aid can be effective in reducing poverty, but only in strong institutional settings and when governments implement “good” policies. It argues that rather than forcing govern-ments to change their policies using policy conditionality, donors should instead target poor countries that already have relatively good governance records and that are following favoured policies.

Aid selectivity can be linked to outcomes such as the achievement of poverty reduction targets specifi ed in the Millennium Development Goals (MDGs). This kind of out-come-based conditionality allows aid recipients greater free-dom regarding what policies to pursue, as long as they keep meeting development targets. This use of targets and out-comes is a feature of General Budget Support arrangements, which seek to ensure that countries in receipt of unallocated aid stay on track. It is also a feature of the HIPC-2 initiative (for highly indebted poor countries), which requires prepara-tion of a PRSP and a record of adherence to macroeconomic policies defi ned by the IFIs. The fi nal result, however, as the case of Mozambique suggests (see box 11.2), is more inten-sive donor involvement in policy-making processes.

BOX 11.2: Aid dependence narrows policy options in Mozambique

One aid modality that has been found to be extremely infl uential in Mozambique is General Budget Support. Under this mechanism,

19 donors coordinate their activities with the government through the Group of Donors for Budget Support to Mozambique (known as

G19). The group holds regular meetings to discuss progress in the country, measured using a performance assessment framework,

which identifi es targets from Mozambique’s PRSP. This outcome-based conditionality is closely aligned with the MDGs and includes

targets on universal primary education, provision of basic health services and success in fi ghting HIV/AIDS.

Mozambique has more donors supplying General Budget Support than any other country.a In fact, the funds generated through this

channel accounted for 27 per cent of aid to that country in 2004–2006. Not surprisingly, many observers have described the relationship

between donors and the government of Mozambique as highly unequal.

General Budget Support was intended to increase national ownership and limit the role of donors. However, in reality it has meant an

increase in donor involvement, with the entire policy process thrown open to donor infl uence. Communication between donors and

recipients at all levels of government is conducted on a daily basis. This has given donors the ability to infl uence policy from within.b

In addition, donor coordination has increased the power imbalance between donors and the government since all foreign aid rests on

their collective conditions.c

Notes: a De Renzio and Hanlon forthcoming. b De Renzio and Hanlon 2008:17. c Batley 2005; Hodges and Tibana 2004.


292

3. Domestic Constraints to Pursuing Redistributive Policies

New democracies must also contend with a number of domestic constraints to pursuing redistributive policies. The fi rst is the limited nature of industrial transformation and class formation. As table 11.1 shows, the labour force characteristics of low- and middle-income countries dif-fer substantially from those of high-income democracies. Union density (meaning union membership as a percent-age of the labour force) and coverage levels are also lower in low- and middle-income countries than in high-income democracies. A labour market structure that is dominated by agriculture and micro-scale informal enterprises may act as a constraint on the formation of interest group organiza-tions and parties of the type that historically affected public policies for redistribution. Recent trends in informalization and subcontracting in large fi rms (see chapter 9) further complicate the problem.

An additional constraint is the quality of democracy in many newly democratized countries. Even though fi gure 11.1 suggests a rise in the number of elections for all coun-

try groupings, elections in many countries are of question-able quality, and rights are not fully institutionalized or protected. Since 1990, more than 40 countries that have held elections have been consistently classifi ed as anocratic – meaning, according to Polity IV, that they are not fully democratic. Freedom House has also listed, on average, more than 50 countries that are not fully free in its yearly ratings of countries for the period 1990–2008. Many elec-tions return incumbents to power and, especially in least developed countries, parliaments are largely dominated by single parties. Data on African parliaments for 2005 sug-gest that there was only one effective party in 21 countries. In 15 parliaments, the largest party controlled 70 per cent or more seats.16 The low quality of democratic transitions suggests that voters may be unable to pressure leaders to deliver on their promises, since the threat of losing offi ce tends to be weak. Furthermore, the politics of most parties in new democracies are not driven by programmes, mak-ing it diffi cult to place them on a Left-Right axis.17 Gov-ernments rely less on parties for policy guidance and more on the state bureaucracy, bilateral donors and multilateral agencies. Parties are often elite creations for access to state power, rather than instruments for aggregating voter prefer-ences or the welfare interests of organized groups.

TABLE 11.1: Labour force characteristics of low-, medium- and high-income countries

Countrya

% part of the urban formal labour force*

% in agriculture**b

% in industry**

% in services**

% who live in urban areas

(2002)**Union

density***c

Low-income n.a. 51.85 13.27 31.81 33.36 10.94

Middle-income 49.95 19.54 24.96 53.99 59.13 24.37

High-income 82.94 4.68 25.96 65.92 76.09 32.72

Notes: n.a. = not available. aAll countries are grouped according to World Bank classifi cations in 2002: high-income (per capita gross national income of $9,076 or more), middle-income ($736–$9,075) and low-income ($735 or less). b Sectoral data for high-income countries are from 1998–2000, middle-income countries from 1995–2000, and low-income countries from 1995–2000. cUnion membership as a percentage of the non-agricultural labour force, 1995. Sources: *ILO 2003. **World Bank 2004a. ***ILO 1997.


293

If parties are not programmatic and voters do not place a premium on programmes in their voting preferences, it is diffi cult to hold leaders to account, and incumbents have no incentive to translate promises into effective poli-cies. The outcome is likely to be targeted provisioning.18 Since voters may not believe that any party is credible, they may opt for individually targeted welfare or “elec-toral bribes” during election cycles. Targeted welfare may include assistance with jobs, loans, cash and goods for individual voters. While these may benefi t some individu-als, they release leaders from the obligation to provide ser-vices that will benefi t the majority. It is not surprising that new democracies have been associated with large political budget cycles. One study19 fi nds that the fi scal defi cits of 44 African countries where multiparty competitive elections have been held increased by 1.2 per cent of gross domestic product (GDP) during election years. Signifi cantly, it fi nds no cycles in countries with non-competitive elections.

Elections in many countries are of questionable quality, and rights are not fully institutionalized or protected

A third constraining factor is ethnicity. As fi gure 11.2 shows, low- and middle-income countries are much more ethnically diverse than high-income countries. A common assumption is that the development of democracy, support for public goods and welfare development more broadly rest on trust and solidarity, which mainly thrive in societies with common values.20 Observers contend that governments in ethnically diverse societies underprovide public goods and redistribute less for a number of reasons: because cross-cultural exchanges are less common than exchanges among members of the same group; differences exist in group preferences for distinct types of public goods;21 inequalities among groups may encourage individuals from richer groups to opt for private provision or oppose redistribution; and it may be diffi cult to apply social sanctions or punishments across ethnic divides.22 This implies that even if voters are well informed about public policies, parties or politicians can only be credible to one segment of

the voting public at best. And, if parties seek to be credible across ethnic divides, voters may be infl exible, rewarding only parties that are perceived to be led by co-ethnics.

Observers contend that governments in ethnically polarized societies tend to underprovide public goods and redistribute less

The fi rst scenario encourages targeted provisioning and patronage politics. In the second, the vote loses its sig-nifi cance as a sanctioning device against bad performance. Politicians may then perceive the ethnic vote as a vote bank that cannot be claimed by rival politicians or parties from other groups. This may also affect the formation of cross-ethnic interest group organizations and social move-ments or weaken the effectiveness of these groups when they embark on collective action. However, as box 11.3 shows, these kinds of pathological outcomes are more likely to occur in societies where the ethnic structure is polarized, not in countries with high levels of ethnic fragmentation.

FIGURE 11.2: Ethnic, linguistic and religious cleavages by income level

0

2

4

6

Frac

tiona

lizat

ion

inde

x

Low-income Lower-middle Upper-middle High-income

Ethnic Linguistic Religious

Note: The fractionalization index measures ethnic, linguistic and religious cleavages. Source: Sheingate 2008, drawing on Alesina et al. (2003), and World Bank (2008a).


294

BOX 11.3: High levels of ethnic fragmentation may be good for democracy and redistribution

The effects of ethnic diversity on behaviour are complex. In a 15-country UNRISD study on ethnic inequalities and the public sector,

for example, it was found that a high level of fragmentation – meaning that no group was large enough to dominate the public sphere

– encourages good inter-group cooperation.

Ethnic structures assume various forms. They vary from cases in which one ethnicity is overwhelmingly dominant and coexists with

numerous small groups (as in Botswana and Lithuania), to cases where two or three groups or two or three large and relatively equal

groups predominate in a multi-ethnic setting (as in Belgium, Bosnia and Herzegovina, Fiji, Latvia, Malaysia, Nigeria, Switzerland and

Trinidad). A third possibility includes cases in which the ethnic structure is very diverse and may take two forms: cases of high levels

of fragmentation (as in Papua Guinea and the United Republic of Tanzania), and cases where fragmentation is interspersed with a few

large and relatively equal groups, which may encourage polarization (as in Ghana, India and Kenya).

The fi ndings suggest that in cases where one ethnicity is overwhelmingly dominant and co-exists with many small groups, ethnicity is

less likely to be important in determining how the public sector is governed. Electoral competitiveness may open up confl icts within

the dominant ethnicity, allowing individuals from minority groups to participate actively in parties led by individuals from dominant

groups. In Botswana, even though most Tswana vote for the ruling party, a sizeable percentage also support opposition parties; and

minorities have not formed separate parties from those led by Tswana. Candidates from the three dominant Tswana subgroups and

the second largest group, the Kalanga, have occupied more than two-thirds of the cumulative parliamentary seats of the main political

parties since 1965. Although the Tswana constitute 70 per cent of the population, they accounted for only 58 per cent of cabinet posts

in 2000, 50 per cent of top civil service posts in 2003, and 61 per cent of parliamentary seats in 2000, suggesting strong representation

of minorities in the public sector. Voting patterns have assumed an urban-rural divide, with the opposition winning most urban votes

and the government winning rural ones. A multi-ethnic elite pact at independence that granted the Tswana language offi cial status

(along with English) was made in exchange for equal distribution of resources among all groups. Only the small pastoral San group

feels strongly excluded.

In highly fragmented settings with relatively small ethnic groups, as in the United Republic of Tanzania, ethnic-based behaviour

in the public domain is also likely to be less virulent, since it is diffi cult for a single group to be hegemonic under free, fair and

competitive conditions. Since political parties may have to appeal to a cross-section of groups to be electorally viable, they are bound

to be multi-ethnic. No single ethnic group dominates the United Republic of Tanzania’s public sector in terms of their share of posts.

The largest ethnic group (which makes up 13 per cent of the population) and the second largest group (4 per cent of the population)

were not represented at all at the top layer of the civil service in 2006. This is not surprising, since it is rare for an ethnic group to have

more than one member at the top cadre of the civil service and cabinet. In the 1990 cabinet, for instance, the largest group had only

7 per cent of posts, the second largest had none, and the third largest (4 per cent of the population) had 4 per cent. A similar distribution

was found in the 1995 and 2000 cabinets, except that the second largest group increased its share from 0 to 4 per cent. The third

largest group was not included in the 1995 cabinet, but got 7 per cent of the total in 2000.

The more diffi cult cases are countries with bipolar and tripolar ethnic structures or cases where ethnic fragmentation is interspersed

with a few large and relatively equal groups that may form selective ethnic coalitions. Countries with these types of ethnic structures

that are relatively stable and cohesive have introduced institutions and policies that are sensitive to ethnicity in order to infl uence the

way in which the public sector is governed.

Source: Bangura 2006.


295

4. When Democracies Deliver Outcomes That Are Benefi cial to the Poor

Some democracies have been able to overcome these con-straints and deliver outcomes that are benefi cial to the poor. This section discusses fi ve broad types of cases involv-ing interest group and social movement activism. It makes four arguments:

Democracies deliver outcomes that are benefi cial • to the poor when groups with strong ties to the poor demonstrate the capacity for organization and mobilization, transcend or reconcile horizontal divisions, and create structural links with actors involved in policy making. This leads, at times, to social pacts.Some success can be achieved without formal • group ties to state actors, but these require high levels of contestation and continuous mobilization to sustain gains. Contestation may block unpopular policies or may lead to the adoption of some popular demands in public policy. However, institutionalization can be a problem without continuous mobilization.Electoral competitiveness in which there is a high • probability that governments may lose offi ce can act as an incentive for redistribution and progressive reforms. However, electoral competitiveness without effective group organization and contestation may produce weak redistributive outcomes.The poor suffer when interest groups and social • movements are weak and when the electoral system is not suffi ciently competitive.

In high-income democracies, organized

interest groups spurred redistribution

In high-income democracies, contestation by organized interest groups produced institutional regimes that allowed groups to bargain with state authorities and infl uence the direction of public policies. In these countries, democracy

and welfare development were driven by similar processes in which trade unions, acting through social democratic and clerical parties, played a substantial role. Socialist/labour voting in the early period of democratization correlated highly with welfare programme consolidation (meaning that countries adopted at least three of the four main social insurance programmes relating to work accidents, health, pension and unemployment).23 However, labour needed allies, since it was unable to effect democratic change and welfare development on its own. Though workers were in the majority in all countries, working-class parties could not obtain electoral majorities. Socialist parties’ share of the vote averaged 30 per cent.24

In the Nordic countries, workers collaborated with seg-ments of the middle class and small farmers, who defended their interests through agrarian parties. The preferences of agrarian parties for fl at-rate, universal, tax-fi nanced ben-efi ts came to defi ne the welfare policies of social demo-cratic regimes. Despite the importance of small farmers in democratization and social policy development, early agrarian democracies or those with no strong labour par-ticipation (such as Canada, France, New Zealand, Norway, Switzerland and northern and western parts of the United States) were laggards in welfare development. In Belgium and the Netherlands, unions and working-class organiza-tions worked through clerical parties whose ideology was critical of free-market capitalism. However, while social democratic parties provided platforms for women’s groups to mobilize for the incorporation of gender issues in welfare policies,25 clerical parties were not so accommodating to gender interests. In the Netherlands, for instance, women perform poorly on labour force participation, unpaid care work and the provision of childcare facilities.26

The bargaining regime of countries with superior social outcomes took the form of social pacts. The key features of such pacts included the recognition granted to representa-tives of labour and employers in negotiations over wages, employment, working conditions and welfare; the ability of group representatives to ensure members’ compliance when decisions were reached; and the mutual recognition of each actor’s importance in achieving goals, including the


296

relative capacity of parties to obstruct outcomes that were not based on consensus. Social pacts were not confi ned to the industrial sector. Agrarian pacts were also constructed

in many countries. These improved farm incomes and narrowed rural-urban inequalities in many countries where farmers’ votes were important (see box 11.4).

BOX 11.4: Social pacts between farmers and the state improve well-being

Historically in industrialized democracies, close links among farmers, politicians and bureaucrats were a key feature in the evolution

of agricultural policies that insulated farmers from the vagaries of market forces. These policies took the form of direct cash transfers,

minimum support prices and various other subsidies. They also included policies that increased opportunities for non-agricultural

employment and environmental policies that provided transfers in connection with conservation or environmental protection measures.

The agricultural welfare state depended on corporatist, or quasi-corporatist, arrangements between producers associations and

agricultural bureaucracies. Often this took the form of a single peak association that enjoyed a representational monopoly in the

policy process, although not necessarily so. In the United States, for example, commodity-based organizations carry greater infl uence

in policy discussions than broader farmers organizations. In many European countries, multiple producers associations organized

along regional, religious or economic lines represent farmers in policy discussions. In political terms, farmers eventually became

important electoral constituencies. Sometimes, these political ties linked farm organizations to parties of the Centre Right, as in

Japan and France. Other times, farmers organizations became core constituencies of the Centre-Left, as in many Nordic countries.

Although the specifi c policies, organizations and political components varied, the agricultural welfare state provided farmers with a

steady stream of benefi ts, access to policy and attention from politicians. Today, in fact, these close relations between farmers and the

state are a major obstacle to reaching political agreement on the reduction of agricultural subsidies in wealthy countries.

In the developing world, relations between farmers and the state look very different. Not many examples can be found of political

incorporation of producers or stable arrangements of policy coordination built around agricultural associations. However, exceptions do

exist, which suggest the conditions under which agrarian social pacts might emerge. In parts of India, for example, policies designed to

increase dairy farming helped to promote producers cooperatives. The programme has been a success, contributing to a pronounced

increase in per capita milk production over the past decades. But there are also indications that cooperative members, many of them

women, have become politically infl uential as a source of votes in local, state and national elections. In Eastern Europe, agrarian political

parties have been infl uential players in the post-communist transition in several countries. In Poland, where 20 per cent of the labour force

is employed in farming, agrarian parties have competed successfully for the rural vote, frequently earning a place as a junior coalition

partner in the government. Finally, in Senegal, peasant organizations, producers cooperatives and other farmers associations came

together in the early 1990s to create the Conseil national de concertation et de coopération rurale (CNCR). With a membership that reached

more than 3 million farmers by 1995, the CNCR plays a central role in the elaboration of government policies towards agriculture.

How do such relationships take shape? Historical legacies matter a great deal, especially with regard to the character of agricultural

policies, the development of rural organizations, and the structure of macro-political institutions. In addition, however, relationships

forged in industrialized countries, as well as those cited in India, Poland and Senegal each occurred in a context of democratic politics.

Although generalizations are diffi cult, it can be argued that competitive elections are conducive to the emergence of agrarian social

pacts. This is because the capacity for agricultural associations to shape sectoral policies depends on the political value of farmers

as an infl uential source of votes. However, an important caveat must be noted as well: farmers do not always or necessarily vote as

farmers. Where regional, ethnic or religious cleavages are the basis of political identities, it will be more diffi cult for producers to

become infl uential in politics and, accordingly, infl uence policy.

Source: Sheingate 2008.


297

High levels of unionization and coverage rates in social democratic regimes encouraged unions to support poli-cies that reconciled wage and welfare demands with the goals of profi tability and growth. Unions supported policies of wage compression and equal pay for equal work across sectors, which spurred employers to raise labour productivity and avoid the option of cheap labour and segmentation.27 Unions were also able to restrain the short-term interests of their members because of their encompassing position in the econ-omy: deals that arose from bargaining had wide worker coverage, and bargaining took place at the national, not industry, level. An important social policy innova-tion that took place in Sweden in the 1960s and 1970s was bringing women’s particular interests (publicly supported childcare, employment) into the policy agenda. This may have been facilitated by union preferences for women’s labour versus immigrant labour to solve the labour shortage problem. But it also involved considerable bargaining and brokering by feminists with men in the Social Democratic Party, who were lukewarm towards fem-inism, and with men in the blue-collar trade unions who were even less sympathetic to many of their demands.28 Although sector-level bargaining may encourage union federations to be concerned primarily with the narrow interests of their own members, some countries with such labour market institutions, such as Germany, have been able to coordinate deals across sectors to produce relatively stable economic outcomes.29 The defl ationary policies of independent central banks in periods of wage-push infl a-tion may force unions to moderate their claims. However, these countries score less on welfare outcomes than those with peak-level bargaining institutions and central banks that are relatively accommodating.

As chapter 5 shows, although social transfers have reduced poverty in all high-income democracies, countries classifi ed as social democratic have been more effective in reducing poverty, followed by countries classifi ed as Christian democratic. Those characterized as liberal regimes, which have weak labour movements and pluralis-tic institutions of interest representation in policy making, are the least effective.

In welfare democracies in the South, subaltern

groups were the main catalysts for change

In a few established democracies in developing countries, subaltern interest groups are fairly well organized and are part of broad social movements that have infl uenced the policies of political parties.30 The most well-known cases are Costa Rica, the Indian states of Kerala and West Bengal, and Mauritius. Because these were largely agrar-ian societies when democratic politics were introduced, peasant movements and organizations were much more active in the construction of the alliances that produced welfare-enhancing policies. In Costa Rica and Mauritius, smallholders displayed remarkable organizational abilities because of the absence of powerful land-owning elites. In Kerala and West Bengal, revolutionary parties implemented land reforms during the early stages of democratization.

In Costa Rica, although the dominant coffee elite owned a few large farms, the bulk of its income was derived from processing and external trade, and not from the exploitation of farm labour. There were many unions representing land-less peasants and defending land invasions between 1970 and 1990.31 In Mauritius, although agriculture was dominated by merchant capital and an agrarian bourgeoisie, there was also a large class of small landholders and rural farm workers. In Kerala, highly unequal agrarian land ownership, discrimina-tory caste structures, and growth of a rural proletariat that followed land commercialization in the south spawned a range of peasant movements for land reform, wage increases and social reforms, which enhanced the role of smallholders and the rural proletariat in the political economy.32

In these societies, subaltern groups were suffi ciently organ-ized to infl uence the orientation of politicians without reliance on intermediaries. In Costa Rica, farmers formed the National Association of Coffee Producers to defend smallholders’ interests on prices, taxes, credit and welfare. In Mauritius, the peasantry collaborated with the growing agricultural labour force, which had formed the Mauri-tius Agricultural Labourers’ Association, and urban trade unions, such as the Engineering and Technical Workers’ Union, which waged active campaigns for labour rights,


298

wage increases and improved working conditions. These organizations played a role in the formation of the fi rst nationalist party, the Mauritius Labour Party, which spear-headed social reforms. Signifi cantly, these interest-based organizations and parties bridged the Indian-Creole eth-nic divide that threatened to undermine social peace.33 In Kerala, because of the twin problems of land alienation and social discrimination, the main social movements were the anti-caste reformist movements in the south, and the land reform movements in the more feudalistic north. All three cases involved a dense network of civil society organiza-tions. In Kerala, unionization often extended to informal sector workers. Social pacts involving the participation of government, labour and business were institutionalized in Kerala and Mauritius. This helped to reduce strikes and improve wages and working conditions.

In these societies, subaltern groups were suffi ciently organized to infl uence the orientation of politicians without reliance on intermediaries

The main political parties embraced a discourse of social rights and equity. In Kerala and West Bengal, the main parties of reform were communist parties, which embraced a parliamentary route to transformation and established strong ties with the peasantry and, especially in Kerala, a small working class. In Costa Rica, although the main parties were not strongly linked to social movements, the elite adopted a social democratic orientation. A reformist communist party (later banned), with ties to a small labour force, helped shape the discourse on social rights when the welfare state was established in the 1940s. In Mauritius, all the major parties defi ne themselves as social democratic and consistently regard social rights as acquired rights by citizens.34 The deepening of democracy and extension of social rights eroded clientelistic relationships.35

In Costa Rica, Kerala and Mauritius, parties routinely alter-nate in government. No single party accounts for all the votes of subaltern groups. In Kerala, despite its transforma-tive agenda, the Communist Party has never ruled for two consecutive terms, and competes with the Congress Party for the votes of the peasantry and rural workers. This has forced the Communist Party to retain much of its social movement character – constantly organizing its base, build-ing alliances with new groups, and making demands that address the interests of its base.36 Electoral competition and active citizenship have also forced the Congress Party to imitate the strategies of the Communists by responding to the needs of the electorate. The failure to build perma-nent winning coalitions suggests that competitive elections play a more substantial role in forcing leaders to pursue redistributive policies than in the social democracies of advanced industrialized societies where social democrats governed for long periods through the electoral alliance of workers, small farmers and the middle class. Some schol-ars have argued that power alternation in Kerala explains its superior social outcomes when compared to West Ben-gal, where Communists have been in power for more than 20 years.37

Interest-group politics have tended to be inclusive, rather than sectarian, because groups forged rural-urban alliances that incorporated wide segments of the population. This has been favourable to both growth and redistribution. Although the growth rates of these welfare democracies did not reach the levels of the East Asian developmen-tal states, they were respectable for much of the period of the 1960s–1990s (with growth in Mauritius reaching 6 per cent; in Costa Rica, it averaged 5.3 per cent in 1963–2000, and 7.6 per cent in 1963–1973). This ensured some economic transformation and funding of extensive social programmes, although Kerala relies substantially on overseas remittance income. Rural-urban alliances also facilitated extension of welfare rights to all citizens. Wel-fare protection in Costa Rica initially covered non-skilled and semi-skilled workers, and later those who were not considered poor. Social insurance coverage in that coun-try was eventually extended to informal sector workers and farmers (see chapter 5).38


299

Remarkable strides in social development have been made in all four cases. Costa Rica spends 16 per cent of its GDP on the social sector. And, unlike dualist regimes with highly regressive welfare policies, social expenditure is shared almost equally among education (31 per cent), health care (31 per cent) and pensions (28 per cent). Social policy is oriented towards increased coverage of the population under a unifi ed system (see chapters 5 and 6). Similarly, in Kerala, a large proportion of the population enjoys social protection and has access to food subsidies; the coverage of health care and primary education is universal. In Mauri-tius, there is a universal basic retirement pension, free pri-mary and secondary education, comprehensive free medical care and subsidized basic foodstuffs. The overall effect is that poverty rates have been drastically reduced in these cases, and literacy and life expectancy rates are comparable to those of industrialized countries.

In Costa Rica, the Indian state of Kerala and Mauritius, poverty rates have been drastically reduced, and literacy and life expectancy rates are comparable to those of industrialized countries

In dualist countries leaning towards welfare

democracy, social pacts have been key

In the context of current democratization, a few other coun-tries appear to be leaning towards a social democratic path, especially some of the dualist or highly unequal regimes in Latin America and in South Africa. Because of their high levels of industrialization, unions in these economies are often well organized and have been able to protect mem-bers’ welfare benefi ts, often at the expense of those with weak or no links to the formal labour market.39 This section focuses on Brazil, South Africa and Bolivia.

BrazilDuring Brazil’s fi rst decade of democratic rule from the mid-1980s to the early 1990s, the new leaders were not strongly committed to redistribution. However, under presidents Fernando Henrique Cardoso and Luiz Inácio Lula da Silva, there was a reversal of trends, representing a move towards more broad-based and expansive social programmes.40 The nature of the coalition underpinning the presidency is one of the key explanatory factors for the change. Cardoso mounted an effective alliance of Left-wing and social demo-crats (the Brazilian Social Democratic Party) and the Cen-tre Right (represented by the Liberal Front Party and part of the Brazilian Democratic Movement Party). In turn, Lula represented the rise to power of the Workers’ Party, with strong ties to the industrial working class and a commitment to the cause of redistribution and welfare policies. Civil soci-ety organizations that campaigned for the impeachment of a former president (Fernando Collor de Mello) through the Movement for Ethics in Politics later turned their attention to the issue of poverty, which gained prominence in policy circles as the democratic regime was consolidated.

There appears to be a virtuous cycle linking comprehen-sive social assistance programmes to electoral competition. The programmes with the highest redistributive impact – conditional cash transfers – have generated a politically effi cacious constituency. The comprehensive nature of the programmes refl ects the dynamics of electoral politics in which presidents respond to the demands of the rural poor that were empowered when the vote was universalized in the mid-1980s. While there is an overlap between the benefi ciaries of one of the most important cash trans-fer programmes, Bolsa Familia, and the rural pension pro-gramme, the clientele of Bolsa Familia in 2008 included 11 million families, affecting probably about 18 million voters; the clientele for rural pensions included 8 million benefi ciary households, representing a constituency of some 12 million voters.

Before the introduction of cash transfer schemes, credit for social programmes was mostly claimed by the politi-cal elite at the subnational level. Historically, funding for education, health and, most notably, social assistance was


300

determined by patronage games and clientelistic networks, typically involving local elites. However, the current cash transfer programmes operate in a fairly transparent man-ner, with minimum corruption or patronage by powerful politicians. Under the new dispensation, presidents have a strong incentive for poverty reduction because they have become accountable for social policy for the fi rst time. Because the constituency of presidents is the whole coun-try, they tend to develop a strong interest in programmes that are not focused on narrow groups of the population.41 However, inequalities remain very high, despite recent progress in reducing them, raising serious questions about the extent to which the project of welfare democracy can be advanced.

In South Africa, democratic processes have driven the expansion of social assistance, which has become the main vehicle for addressing the plight of the poor

South AfricaIn South Africa, democratic processes have also driven the expansion of social assistance, which has become the main vehicle for addressing the plight of the poor in the light of a growth strategy that has failed to generate invest-ment and employment.42 In part, this failure was due to the racially repressive labour regime that created extremely high income differentials during the apartheid era. The subsequent struggle for democracy produced an alliance between the main trade union organization, the Congress of South African Trade Unions (COSATU), and the Afri-can National Congress (ANC), which resulted in a joint programme for reconstruction and development. The new ANC government established a multipartite body to reach consensus on economic and social policy among organ-ized labour, organized employers, community groups and

the government. However, consensus on macroeconomic issues proved elusive as the government’s policy shifted in a more liberal direction. Meanwhile, the labour movement pressed relentlessly for improved wages and often backed up its demands with strikes. Employers failed to invest profi ts and expand employment, preferring to transfer capital overseas.

Despite these tensions, the pact between the labour move-ment and the government has survived; when combined with electoral pressures, it has tended to push the gov-ernment in a social democratic direction on social policy. Government spending on social assistance programmes has grown rapidly – from 2 per cent to about 3.5 per cent of GDP between 1994 and 2006. Expenditure doubled in real terms between 1994 and 2004. By mid-2006, about 11 mil-lion grants were being paid each month. One in four South Africans receives a pension or grant fi nanced out of general taxation (see chapters 5 and 8). However, the very high levels of inequality inherited from the apartheid era act as serious constraints on the development of social democracy. The record of social movements in pushing through progres-sive agendas is mixed. For example, social movements such as Treatment Action Campaign have collaborated with COSATU to change the government’s policy of neglecting the AIDS epidemic. However, rural social movements have been unable to affect public policy on land reform, which remains wedded to a market-friendly willing buyer–willing seller framework. Only 3.5 per cent of the area designated as commercial farmland had been redistributed by 2005.

BoliviaBolivia represents a mineral-rich country in which a gov-ernment with strong ties to social movements has attempted to change the country’s welfare trajectory in favour of redistribution. 43 In countries with extractive industries, by far the most signifi cant channels through which they can fi nance programmes for poverty reduction are tax and roy-alty payments to government. The extent to which these contributions are made, and how they are used, has become a particularly confl ictive axis of activity for social move-ments, and one that brings such movements and states closer together. The current governing party (Movimiento


301

al Socialismo, or MAS), grew out of social movement processes, especially among coca producers and the more historic sindicalist highland peasant movements. These origins inevitably gave the movement/party a heavy anti-imperialist ideology. There was a clear resonance between these movement processes and the two major “resource wars” in Bolivia over the last decade (over water in 2000, and over gas between 2003 and 2005), each of which contested the control of strategic national interests by international companies. Not surprisingly, within the fi rst year of its election, MAS passed decrees to nationalize ownership of hydrocarbons and refi neries.

Bolivia represents a mineral-rich country in which a government with strong ties to social movements has attempted to change the country’s welfare trajectory in favour of redistribution

Issues of social protection and targeted support to the elderly and to mothers became an issue for the movement when it came into power (see chapter 8). However, the need to fund programmes such as these generated other challenges, both within submovements of the MAS as well as with opposition movements. Lowland indigenous groups, which generally support MAS, perceive expansion of hydrocarbon production as compromising the security of their livelihoods. The fact that MAS tried to retain some of the revenues from hydro-carbons usually returned to the departments from which they were extracted has led to tensions with the opposition movements. This has renewed dynamism within the older separatist movements among non-indigenous (and politi-cally conservative) populations in the lowland provinces, who complain that this policy change is taking assets away from them. During 2008, movements harnessed this concern and were able to initiate civil disobedience and direct action

on a massive scale that, for a short while, called into ques-tion the viability of the MAS government. Ultimately, the government survived and emerged strengthened. However, the more general point can be made that one movement’s efforts to use the instruments of government to appropriate resources for poverty reduction can – depending on where those resources come from – elicit responses from other movements resisting them. Deeply entrenched inequalities across a number of economic and social dimensions constrain efforts to pursue a unifi ed and progressive social agenda.

Gains are possible, even if interest

groups lack ties to power

Some gains can be achieved even in contexts where con-testation has not produced strong ties with ruling parties or institutionalized regimes of bargaining. Much group activ-ism in the context of top-down democratization is of this type. Although unions were less effective in shaping the ideologies of the main parties in the Republic of Korea and Taiwan Province of China, contestation and electoral com-petitiveness propelled the interests of subaltern groups onto the public agenda. Social movement activism in developing countries often takes the form of direct action by relatively autonomous groups, with positive outcomes for livelihoods in some cases. This section focuses on the cases of the Republic of Korea, Taiwan Province of China and Peru.

Republic of Korea and Taiwan Province of ChinaDemocratization in the successful developmental state of the Republic of Korea and the equally successful Taiwan Province of China was associated with inclusive welfare policies. In both cases, democratization occurred in the context of a major fi nancial crisis that required reform of the labour policy of life-long employment. The rise in unemployment and part-time work generated calls for the expansion of welfare benefi ts. However, despite high lev-els of industrialization, the policies of political parties did not refl ect the classic capital-labour cleavage, and unions were less effective in infl uencing the policies of the main parties. Korean parties were driven by charismatic leaders who eschewed links with labour unions, preferring instead to mobilize votes on the basis of regional calculations and


302

selective social concerns. A social pact involving govern-ment, labour, employers and party leaders was constructed at the height of the 1997 Asian fi nancial crisis and helped rebuild business confi dence and expand social assistance.44 Despite the establishment of a tripartite commission, com-mitment by both government and labour to this mode of interest intermediation has remained weak. Taiwanese par-ties maintained ties with labour unions; however, these were pragmatic rather than ideological, since the basic cleavage was between parties that advocated national independence and those that preferred reunifi cation with the mainland.

The developmental regimes in both the Republic of Korea and Taiwan Province of China were based on wage repression, labour acquiescence and offi cially sanctioned, government-dependent unions, which operated mainly in large enterprises: the chaebols in the Republic of Korea and government-owned enterprises in Taiwan Province of China. This may partly explain why, despite their level of industrial development and democratization, both places have very low levels of unionization (13.6 per cent of the labour force in the Republic of Korea and 10.5 per cent in Taiwan Province of China). The bulk of the labour force is employed in medium and small enterprises (87 per cent in the Republic of Korea and 78 per cent in Taiwan Prov-ince of China) and is not unionized.45 However, patterns of government-labour relations differ. Labour has some ties with political parties in Taiwan Province of China because democratization started at the local government level, where unionized workers’ votes were often decisive. The two main union federations, the Chinese Federation of Labour and the Taiwan Confederation of Trade Unions, are each allied to the two main parties, the Kuomintang Party and the Democratic Progressive Party respectively. Unions have taken advantage of the single non-transferable vote system, in which voters choose candidates in multimember constituencies, to punish candidates, including those sup-ported by their respective labour federation, if they ignore labour demands. Unions’ electoral power at the central level is, however, rather weak.

In the Republic of Korea, diffi culties in consolidating the party system (party splits and mergers are common) have

resulted in less effective links between parties and unions and account for the high levels of labour militancy, includ-ing formation of a labour party by the Radical Union Federation. The relative autonomy of unions and social movements in the country encouraged formation of broad advocacy coalitions for social policy reform, which focused on unemployment insurance; public works, training and basic income schemes that did not require means testing; family policies; and a universal health service system.46 As social conditions deteriorated in the face of the 1997 fi nan-cial crisis, welfare reform became a vote-winning issue, which the opposition party was able to exploit to win offi ce for the fi rst time. In Taiwan Province of China, even though unions and social movements were attached to the main parties, they played virtually no role in setting the agenda for social policy reform.47 The national health system, for instance, was established by the previously authoritarian government, anxious to shore up its legitimacy and win bit-terly contested elections, even though the main advocate of reform was the opposition party.

In both the Republic of Korea and Taiwan Province of China, electoral competition introduced conditions of uncertainty for the ruling authorities and opposition parties and forced them to implement redistributive policies

In both places, electoral competition introduced conditions of uncertainty for the ruling authorities and opposition par-ties and forced them to implement redistributive policies. However, weak union-party links in the Republic of Korea, or lack of a policy stance that refl ects subaltern interests in parties with union links in Taiwan Province of China, act as a limit on the extent to which these now democratic developmental regimes can push the redistributive agenda along more social democratic lines.


303

PeruThe case of Peru48 represents social movement activism in a highly unequal and less industrialized society. Movement activism involves direct action without an institutionalized regime for bargaining. Interactions between social move-ments and the state may be characterized more by recurring cycles of confl ict, which in some contexts have produced pro-poor outcomes. Over the last 15 years, Peruvian gov-ernments have increasingly tied their development strat-egy to a rapid expansion of mineral, oil and gas extraction, affecting over half of the country’s peasant communities and spawning a movement dedicated to the defence of live-lihoods of communities affected by mining. The response to hydrocarbon extraction has also been led by existing indigenous peoples’ organizations under the umbrella of the Association of Indigenous Peoples for the Development of the Peruvian Jungle.

Despite their weaknesses, these movements have made the extractive industry and its links to poverty, the environ-ment and indigenous peoples a topic of public debate. Tech-niques have often been confrontational, and sometimes violent, and have prompted clamp-downs and repression. While justifi cation of such techniques remains question-able, it is also true that neither government nor companies paid attention when direct action was not used. Moreover, the use of direct action has opened political space in which negotiations over policy, though not institutionalized, have occurred. Negotiations have allowed for more protection of the asset bases of local populations, including the intro-duction of water monitoring programmes around mining sites, and greater recognition of land rights. They have also contributed to an increase in revenue transfers to mine-affected areas: in 2004, the government ruled that 50 per cent of the taxes paid by mining companies to central gov-ernment would be returned to the regions of extraction. Negotiations also called into question the adequacy of existing public institutions for ensuring that growth results in poverty reduction – an issue that became prominent in the 2005–2006 presidential election campaigns. In each of these instances, however, these changes have come about because other actors (including those from industry) have also come to support them, partly as a result of protests.

In countries with weak civil societies,

electoral competition can bring leaders

to account

The structure of labour markets in agrarian economies can act as a constraint on interest-group pressure for wel-fare development, although, as an earlier section of this chapter showed, these constraints have been overcome in some cases through well-organized civil societies and broad-based parties. In many low-income agrarian socie-ties, NGOs and donors play active roles in poverty reduc-tion. However, their capacity to effect real change has been limited because of weak links with broader groups in society. In these types of democracies, electoral competi-tiveness can be an important instrument for extracting accountability from leaders. One study49 provides evidence for the ways in which democracy has affected government priorities and shifted pro-urban biases in educational provi-sion in Africa. Democratization made governments more responsive to the needs of the rural poor, who constitute the majority of voters. However, as section 3 has shown, the electoral system in most agrarian democracies is not competitive enough. Governments enjoy huge parlia-mentary majorities and retain the capacity to immunize themselves from electoral defeat. The lack of electoral competitiveness and low density and strength of associa-tions and movements often make it diffi cult to sustain gains outside of electoral cycles.

In many low-income agrarian societies, despite the limited nature of the wage economy, unions have, on numerous occasions, been able to pressure governments and employers to act on the livelihoods of workers

Despite the limited nature of the wage economy, unions have, on numerous occasions, been able to pressure govern-ments and employers to act on the livelihoods of workers.


304

In alliance with other groups, they have also managed to resist policies such as price increases on basic commodities and services.50 However, unions’ resistance to adjustment programmes exposed them to attack from reforming gov-ernments and their international backers, who argued that rural poverty was a product of discriminatory trade and pric-ing policies favoured by an urban coalition that included the working class. Authoritarian measures were supported to free markets and release the grip of the so-called urban coalition on public policy.51 While these measures weakened union power and undermined workers’ livelihoods, they did not necessarily improve the incomes and power of farmers.

Despite their dwindling membership base, unions played an important role in the process of democratization because of their strategic location in the modern economy, although links with governing parties remain tenuous

Organizations of subaltern groups participated in the wave of democratization that swept through countries in the 1990s, although their capacity to subsequently infl uence the direction of policy has been limited. The role of unions in democratization was particularly signifi cant because of their strategic location in the modern economy, despite their dwindling membership base.52 In some African cases, such as Zambia and Zimbabwe, unions provided leadership for a coalition of forces, although the links with govern-ing parties have remained tenuous since elites and external actors dominate the reform agenda. In other cases, such as Ghana, unions have avoided formal ties with parties, preferring to exert pressure as independent actors. Such autonomy does not always guarantee success, since Ghana-ian unions’ hold on public policy remains weak.53 In Sen-egal, a plurality of trade union federations are affi liated to different parties, although their infl uence on party policies is limited. In many other agrarian settings, such as Uganda and the United Republic of Tanzania, where the industrial

sector is small and crisis and adjustment have taken a toll on union membership, governments have ignored union pressures for redistributive policies.

Despite these constraints, there is evidence of gains at the micro and sectoral levels for groups in both urban and rural areas. The most notable has been the dramatic growth of producer organizations in rural settings. In Burkina Faso, the number of villages with at least one rural producer organization increased from 22 per cent to 91 per cent between 1982 and 2002; local associations increased in Senegal from 10 per cent to 65 per cent of villages over the same period.54 There is also evidence of rapid increases in the number of producers organizations in parts of Benin, Cameroon, Chad, Ghana, Kenya and the United Republic of Tanzania.55 Part of this growth can be attributed to international donors that helped promote such organizations in the 1990s. However, the effect of these organizations on poverty reduction is ambiguous.56 It seems, however, that they have had some impact in countries where elections have been fi ercely and fairly contested, such as Ghana and Senegal.57

In the midst of Senegal’s economic crisis in the 1970s, a number of peasant associations began to develop. Encour-aged by a former extension service offi cial who was presi-dent of a local NGO, and with support from international donors, a dozen associations formed a national federation in 1976. When SAPs ended most government interven-tions in the countryside, the associations provided agri-cultural services to their members. Membership expanded to 24 associations of 2,000 village groups representing over 400,000 members. Taking advantage of a national election in 1993, the federation organized a forum that included government offi cials and donors.58 Seven organi-zations that had participated in the forum later created the Conseil national de concertation et de coopération rurale (National Council for Rural Dialogue and Cooperation), or CNCR. The Council became “a single, authoritative interlocutor . . . a full participant in a range of programs and policy forums”.59 The CNCR’s decision to walk out of negotiations that had stalled over the Agricultural Struc-tural Investment Programme in the mid-1990s forced the


305

country’s president, who was worried about the rural vote, to convene a meeting with 150 CNCR representatives. The meeting resulted in an agreement to cut agricultural credit interest rates, remove import taxes on agricultural inputs, issue a fi ve-year moratorium on farmers’ debts, and institute a series of regular meetings between CNCR and the agriculture ministry.

5. The Politics of Poverty Reduction: Implications for Policy

Power relations lie at the core of development. They deter-mine which actors are in a position to fashion and defi ne policies, infl uence development processes, govern the actions of others and participate effectively in economic, social and political life. In this sense, power is important in understanding all social outcomes. Strategies that seek to bring about changes in poverty and inequality must also consider ways to shift the balance of power. Active citizen-ship or group activism is important in generating such shifts and ensuring that governments respond to social needs.60

The driving forces that push democracies to deliver redis-tributive outcomes are complex. Although democracy has been embraced as a core value of development, there is lit-tle understanding of the politics and institutions that enable democracies to promote growth and reduce poverty. The strong wave of democratization that swept much of the devel-oping world in the 1990s and enjoyed strong donor support occurred simultaneously with donor embrace of technocratic styles of policy making. This tended to create what have been described as choiceless democracies.61 Allowing economic policy to be formulated by small groups of experts is not only a refl ection of the centrality of the economic problem, but also a deliberate attempt to get politics out of economic policy. It is informed by the belief that policy is best designed by insulated technocracies. And it is here that there is an obvious and serious contradiction between the politics of insulating technocracies and calls for greater accountability

and transparency in public affairs. Technocratic govern-ance severely limits the development of a culture of dialogue and compromise by foreclosing debates on a wide range of issues that are relevant to poverty reduction, such as income distribution, taxation, employment expansion, and protec-tion or non-protection of certain economic activities. Any economic programme has its trade-offs among which politi-cal choices and decisions must be made. Outcomes of such debates cannot be fi xed a priori in democracies.

The introduction of PRSPs in low-income, aid-dependent countries represented an effort to transfer ownership of economic policies to national governments and democra-tize policy making through involvement of civic groups in designing and implementing anti-poverty strategies. How-ever, participation in PRSP forums has taken the form of consultation without power to effect real change or to get policy makers to deliver on agreed goals. The macroeco-nomic issues that have a bearing on the livelihoods of the poor are not open to substantive discussion, and citizen groups often feel that decisions on important policies are made elsewhere.

Strategies that seek to bring about changes in poverty and inequality must also consider ways to shift the balance of power

Improving the bargaining regime for poverty reduction requires interventions on several fronts.

Abandon technocratic styles of policy making

The technocratic styles of policy making tend to eschew contestation and favour infl ation control over the equally important goals of employment and welfare protection. In order to pursue strategies that are growth-oriented, employ-ment-centred and redistributive, governments and citizen groups need a certain degree of autonomy and the policy


306

space in which to consider a broad range of approaches. The global fi nancial and economic crisis, which has exposed the destabilizing tendencies of fi nance capital and de-legitimized neoliberal policies, offers opportunities for alternative styles of policy making. However, their outcome will depend on the resources countries are able to mobilize to pursue alternative policies and the balance of forces that will inform the content of those policies.

Go beyond NGO–centred PRSP forums

and more actively engage associations

and social movements

It is important to engage more actively with associations and social movements whose members’ livelihoods are affected by government strategies and global change. This may require multiple pacts because of the fragmentation of interests associated with globalization. As already noted, social pacts are often needed to effect deep and long-term changes in the welfare of low-income groups, and have been formed in both industrial and agrarian settings.

Institutionalize rights to encourage effective

group participation in policy making

As this chapter has shown, many new democracies are still fragile, civil and political rights are not fully institutional-ized, and incumbents often have the capacity to frustrate the choices of voters by rigging elections. The institution-alization of rights will empower groups to exercise political choice and hold leaders accountable.

Support organizations that represent the

views of diverse groups that include the

participation of formal and informal workers

The existence of organizations that support the interests of producers is vital to success. Unorganized farmers, for exam-ple, are often forced to compete with one another to sell more in harvest time, which may lead to a fall in prices. A central

organization that represents their interests is able to moni-tor a large number of individual producers and draw up and enforce production and marketing agreements. Similarly, in an environment without unions, competition among indi-vidual workers may lead to lower wages and benefi ts. Peak organizations and coalitions enable groups to overcome collective action problems of free-riding or defection, and strengthen capacities to moderate claims or meet targets. They also help reconcile redistribution and growth. All par-ties in such a bargaining regime must have strategic goals and feel a need for cooperation.

Recognize the fact that not all groups can

be incorporated into bargaining regimes

Accountability and progressive outcomes also require groups with suffi cient clout and independence to engage the state and raise critical issues without the constraints of institutionalized pacts. Building expertise and organiza-tional capacities is vital for group action that can promote progressive change in and out of social pacts. Taking inter-est associations and movements seriously in policy making may spur groups to develop institutional capacities and act responsibly – even in the context of limited resources.

Empower the general population to exercise

infl uence on how polices are made

Empowering the general population to infl uence policies can often produce good anti-poverty outcomes. Conversely, policies that lead to a reduction in poverty and inequal-ity improve social solidarity (a cornerstone of citizenship), lock in disadvantaged groups to the democratic regime by undermining violent alternatives, weaken clientelist social relations, and enhance the capacity of citizens to participate in public life as autonomous actors. In this way, democracy and poverty reduction can become mutually reinforcing.


307

Notes

1 Sen 1999.

2 Meltzer and Richard 1981; Melo 2008.

3 Therborn 1977.

4 Acemoglu and Robinson 1999; Boix 2003.

5 Stephens 2007.

6 Bebbington 2009.

7 Lipset and Rokkan 1967.

8 Schmitter 2006.

9 Segura-Urbiego 2007.

10 See the contrasting results of Segura-Urbiego (2007); Kaufman and

Haggard (2008); Avelino et al. (2005); and Ross (2006).

11 UNRISD 2004a; Corrales 2004; Montecinos 2004;

Ágh et al. 2005; Mansfeldovà 2005; Chirwa and Munthali 2002;

Akindes and Topanou 2005; Mkandawire 2004.

12 Corrales 2004.

13 This section draws on Lavers (2008).

14 Molenaers 2008.

15 Dollar and Pritchett 1998.

16 Bangura 2005.

17 Manning 2005.

18 Keefer and Khemani 2003; Robinson and Verdier 2002.

19 Block 2002.

20 Miller 1995; Miguel 2004.

21 Alesina et al. 1999.

22 Miguel 2004.

23 Hicks 1999.

24 Stephens 2007.

25 Huber and Stephens 2001.

26 Sainsbury 1996.

27 Huber and Stephens 2001.

28 Hobson 2006.

29 Iversen et al. 2000.

30 Sandbrook et al. 2007.

31 Cortes and León 2008.


33 Bunwaree 2007.

34 Bunwaree 2007.



37 Heller 2005; Harriss 2005; Mukhopadhyay 2000;

Sandbrook et al. 2007; Chibber 2008.

38 Martínez Franzoni 2007.

39 Seekings and Nattrass 2008; Melo 2007a.

40 This section draws on Melo (2007a).

41 Melo 2007a.

42 Seekings and Nattrass 2008.

43 This section is based on Bebbington (2009).

44 Choi 2000.

45 Lee 2005.

46 Kwon 2005; Peng 2009.

47 Wong 2005a.

48 This section is based on Bebbington (2009).

49 Stasavage 2005.

50 Seddon and Zeilig 2005; Beckman and Sachikonye 2002;

Atieno 2008; Olukoshi 1998b.

51 Beckman 1992; Toye 1992.

52 Beckman and Sachikonye 2009.

53 Beckman and Sachikonye 2009; Akwetey 2001.

54 Bernard et al. 2008:10.

55 Mercoiret 2006:3, 4.

56 Bernard et al. 2008:6.

57 Sheingate 2008.

58 McKeon 1999:3,4.

59 McKeon 1999:4.

60 Chronic Poverty Research Centre 2009; Green 2008; Kohli 2008;

Wignaraja and Sirivardana 2005.

61 Mkandawire 2006.

CHAPTER Democracy and the Politics 11 of Poverty Reduction

Documents